Parenting Machines: AI teaches students how to resolve conflicts

Students receiving education in the field of hospitality have begun to learn how to solve conflict situations using AI. So, in the first and second courses of the RANEPA, they tested an interactive simulator — students had a conversation with a virtual client in English, working out, for example, hotel check-in situations and complaints about unsanitary conditions in the room. The AI not only conducted a dialogue, sometimes in a very aggressive form, imitating a hotel guest, but also assessed the professionalism of the participant's actions, as well as his language skills. As a result, students who completed the course with AI demonstrated more diverse and successful conflict resolution options than their colleagues who studied traditional methods. See the Izvestia article about other areas in which simulators based on generative AI can be useful.

How AI teaches students

First- and second-year students of the RANEPA Faculty of Hospitality and Sports Industry Management became participants in an educational experiment. During the training, they used an interactive simulator based on generative artificial intelligence (a type of AI that creates, not just analyzes, information): This is how students trained to resolve conflicts and problematic situations that may arise with hotel guests. Izvestia has a report on the results of the study at its disposal.

The experimental group first took a course on "Conflict resolution in a restaurant and a hotel" in English, and then consolidated it using an interactive simulator. The students acted as managers of hotels and restaurants, and the neural network imitated the behavior of a disgruntled guest. At the same time, the simulator automatically analyzed and evaluated both the participant's professional actions and language skills (grammar, professional vocabulary).

Students independently completed sessions with the simulator, repeating them to improve the result. At the end of each of them, the system generated personalized tasks and recommendations for improving communication and linguistic competencies. The neural network also provided a conversation plan and an analysis algorithm.

For example, when a conflict situation was being worked out when checking into a hotel, in which a student offered a "guest" a free room upgrade, the simulator stated that the client received "more than he justly deserved." But the main complaint against the student was that he "did not take advantage of the moment to create loyalty."

"It was necessary to emphasize the exclusivity of such a gesture with the phrase: "We usually do not make upgrades in such cases, but we will make an exception for you," and hand over a business card with the manager's personal contacts with the words: "If something else bothers you, call me personally," the AI simulator advised.

In another scenario, the system recommended not to immediately meet the guest halfway, but to explain the possible motives of his behavior in order to prevent manipulation. And evaluating the conversation, the simulator offered a variety of conflict de-escalation techniques.

In the control group, students practiced similar role-playing scenarios in the classroom, performing exercises without using an interactive simulator. The teacher evaluated their actions according to the same criteria that were embedded in the neural network.

What results did the experiment show?

The students of the experimental group demonstrated results 70% higher than those shown by the control group. In addition, 68% of the participants in the experiment who worked with the simulator improved their English language proficiency.

According to the authors of the study, such results indicate a significant potential for using the interactive simulator in the educational process. In their opinion, this method is a promising means of teaching in the field of socio—humanitarian sciences, as it combines the advantages of the digital educational environment and the methodological capabilities of artificial intelligence.

"The use of the interactive simulator creates a synergistic effect through a combination of realistic feedback and an individualized approach, which contributes to the accelerated development of communication skills, professional terminology and students' confidence in conflict resolution in the hospitality industry," the study says. "Unlike a teacher who is limited in time and resources, the simulator is able to generate an almost unlimited number of unique dialogue scenarios, each time simulating new combinations of complaints, emotional reactions of the guest and professional contexts."

The study notes that in the future it is planned to expand the sample to include students from other specialized programs in the experiment.

Where else can interactive simulators be used?

One of the authors of the study, Peter Ototsky, project director of the Directorate of Priority Educational Initiatives, noted that an AI simulator is currently being tested at the Novgorod branch of the Presidential Academy (RANEPA) to develop the client-centric skills of MFC employees as part of the additional professional education program.

"There is a great potential for using the AI simulator to learn foreign languages," he said.

And ITMO University has developed "critical tools" based on AI to work out the submission of grant applications, the content of educational programs and the training of thesis defense, Alexander Bukhanovsky, head of the Strong AI in Industry research center at this university, told Izvestia.

— This is the use of generative AI to train professional communication skills — imitation of an "evil expert": opponent, reviewer, critic, — he said. — The basis for his work is any creative object created by man. And AI is a harsh critic, formulating shortcomings based on a set system of criteria. At the same time, some of the disadvantages may be specially invented. This is necessary so that a person does not get lost in the flow of criticism and can, without panicking, objectively evaluate the opponent's speech and respond to it in a reasoned manner.

According to Alexander Bukhanovsky, in addition to developing skills in the hospitality and retail sectors, such systems can also be used in the museum sector — when a visitor is offered a personal AI interlocutor who, for example, discusses the museum exhibit with him, makes judgments and accepts criticism from a person.

Konstantin Bakulev, Deputy head of the Department of Engineering Cybernetics at NUST MISIS, told Izvestia that such simulators have a "great future", including testing the development of communication skills among administrators, teachers, doctors and other representatives of socially oriented professions.

"The main thing is that this interaction between a machine and a person should be monitored and tested in one way or another by a special maintenance team that would monitor and remove, including through retraining of the model, accidental errors and hallucinations," the expert said.

Artem Egorov, head of the laboratory of Immersive Technologies in Education at the Skolkovo School of Management, said that the use of AI simulators to practice communication skills is becoming a key tool for solving problems in any field where success depends on the quality of dialogue.

— On the one hand, this is a massive training of personnel working with clients— such as call center operators and sales managers. Here, AI allows you to hone scripts in a secure environment, work with objections, and de-escalate conflicts. On the other hand, this is a more complex level: simulators are used to develop managers," the expert said.

In addition, for managers, such a system provides an opportunity to practice providing constructive feedback, resolving disputes within the team, or managing employees with low productivity.

— The key value of technology is not just in simulation, but in instant and objective feedback. AI analyzes not only what is said, but also how: the pace of speech, the confidence of tone, the manifestation of empathy. It acts as an accelerator of communication practice, allowing an employee to complete a path in weeks that previously took years," he said.

In his opinion, simulators can also help prepare doctors to report complex diagnoses, educators to resolve conflicts with parents, and lawyers to negotiate.

— Probably, the discussion will soon shift from the question "Is it effective?" to the question "How to integrate AI simulators into corporate and professional training?" — said Artem Egorov.

However, the IT company MWS AI emphasized that the effectiveness of performing certain actions in a particular scenario depends on how well a person has set the framework for such a simulator.

If there are articles and literature that can help a person resolve a conflict situation, it means that AI can be trained to work with this scenario at a basic level, experts noted.